On President Trump and Popular Mandates

President Trump, in his first week in office, appears to be concerned that he be perceived as having a popular mandate. This is my explanation for his doubling down on the unfounded claim that he would have won the popular vote if it hadn’t been for 3-5 million (!) illegal immigrants voting. Much has been written about how false this claim is, but I won’t waste time on that here.

Both Trump and Clinton failed to achieve 50% of the popular vote, at 62.99 million votes (45.94% of votes cast) and 65.85 million votes (48.03%) respectively. Gary Johnson, the Libertarian candidate, was third place, with 4.49 million votes (3.27%). Jill Stein, the Green candidate, was fourth with 1.46 million votes (1.06%). On top of that, Evan McMullin, an Independent from Utah, won 0.73 million votes (0.53%), and managed to win over 20% of Utah’s popular vote. Indeed, as I have shown in another post, if the states allocated their Electoral College votes fairly and efficiently, the final EC vote tally would have been 261 (Trump)-261 (Clinton)-14 (Johnson)-1 (Stein)-1 (McMullin). I find that interesting, but it doesn’t point to a mandate for anyone. Let us suppose that:

  1. Voters had been allowed to rank their choices.
  2. The actual 2016 results actually reflect the #1 choices that voters would have made.
  3. Since no candidate hit the 270 EC vote mark in the first tally, then there is an automatic runoff. In national rounds, states eliminate the bottom EC vote recipients one at a time, in order from least popular votes (in that state) to most. In each round, they re-allocate their EC votes adding in the lower ranked choices of the ballots where the higher ranked choices have been eliminated. This process iterates until there is a candidate with 270+ EC votes, or there are only 2 candidates with EC votes left. (If my explanation is confusing, watch this short video that explains how single-winner instant runoff elections work.)
  4. Voters who put Stein as their #1 choice may safely be assumed to mostly have put Clinton as their #2 choice.
  5. Voters who put McMullin at #1 may safely be assumed to mostly have put Trump or Johnson as their #2 choice.
  6. Enough voters who put Johnson as their #1 or #2 choice, put Trump as their next choice, which puts him over the top with 270+ EC votes. (A 269-269 draw is also possible, which would punt the choice to the House of Representatives. Here, though, I want to assume a Trump victory.)

In this ranked choice voting (a.k.a. instant runoff) scenario, the President would have both his Electoral victory and a popular vote victory of sorts. Indeed, I imagine much of the current rancor from the Left wouldn’t exist, because there would be more of a feeling that everyone had been represented in the choice. In fact, Trump could look at the instant runoff details, see for himself that up to 4% of his win came from voters who preferred Johnson or McMullin, and adjust his rhetoric to say that he understands his mandate also came from them.

Now, forget my fantasy scenario above, and look again at the actual popular results. Unfortunately, they are much less information rich about the will of the electorate. Even so, Presidents who don’t secure 50% of the popular vote will normally take a look at those “spoiler” votes. They then consider how they might also appeal to and be the president for those voters as well. Indeed, ideally, they should also consider how to have constructive dialog with their major opposition.

Or they could just do as the current President is doing, and blame any problems on certain foreign groups and illegal immigrants.

Added Fair, Efficient Electoral Tally for 2000

I’ve been wanting to do this one, the other election in recent memory where the national popular vote winner lost in the Electoral College (EC). I’m, of course, talking about Bush v. Gore in 2000. I updated my GitHub project, adding a notebook that scrapes the 2000 state-wise results from Wikipedia, and applies my fair, efficient EC tally method. Here’s what I found:

Candidate Party Electoral College Votes Wasted Popular Votes
George W. Bush Republican 263 765,498
Al Gore Democratic 262 863,577
Ralph Nader Green 13 1,197,844
Pat Buchanan Reform 0 448,895
Harry Brown Libertarian 0 384,431
Howard Phillips Constitution 0 98,020
John Hagelin Natural Law 0 83,714
Others 0 51,186
U.S. Total 538 3,893,165

The all-important Florida count would have been a non-story, with no need for the Florida and U.S. Supreme Courts to weigh in. The EC count from Florida would have been Bush 12, Gore 12, and Nader 1, with it being not even close on allocating that last EC vote to Nader. Bush and Gore each had about 50k remaining popular votes after getting their 12 EC votes, to Nader’s 97.5k votes.

Just as in 2016, the real story is that both candidates failed to win a clear majority of the popular vote, with substantial voting for so-called 3rd party candidates. However, in 2000, the popular vote difference between the leaders was much smaller. The baked-in EC advantage for less populous states shows in the fact that the national popular vote loser still got one more EC vote than the national vote winner, even using my method. Recall that my method tries to make the allocation fair and efficient only on the state level, since it’s the states that pick the EC members. Also, in 2000, it is clear that Nader and the Green party caused an election upset for the Democrats, since there is a lot of philosophical similarity between the Green and Democratic parties.

In 2000, the failure of the Democrats to secure the White House can be pinned on voters that chose Ralph Nader and the “winner take all” allocation method. This is especially true looking at Florida, where all those Nader voters would probably have preferred a Gore victory to what happened. (Of course, as I’ve stated before, I would prefer the EC members be allowed to deliberate when no candidate gets the requisite 270 votes in the first round. This would give the 3rd party EC members a chance to caucus and influence. In 2000, this would likely have meant an ultimate 275-263 Gore victory.) By contrast, in 2016, most of the 14 Libertarian EC members would likely have caucused with the 261 Republican EC members, still resulting in a Trump victory.